GitHub - openai/gpt-2: Code for the paper "Language Models are Unsupervised Multitask Learners"

▤ Summary

OpenAI shares the GPT-2 code and models from the paper 'Language Models are Unsupervised Multitask Learners.' The repository serves as a starting point for researchers and engineers who want to experiment with the model. The text warns that GPT-2 may be biased, inaccurate, or unsafe in some use cases, especially without fine-tuning. It also notes that OpenAI released a dataset for research and may release evaluation code and larger models later. The post invites researchers working on synthetic text detection, bias mitigation, and other GPT-2-related studies to reach out.

▦ Linked Companies

OpenAI

$25M to $50M